Similarity-based competition in relative clause production and comprehension

• We investigate similarity-based interference in production and comprehension.
• Conceptual similarity influenced planning and comprehension of relative clauses.
• Both tasks show similarity effects at verb phrase processing.
• Competition in event role-to-sentence mapping may be shared across tasks.

This work investigates the role of semantic similarity in sentence production and comprehension. Previous research suggests that animacy and conceptual similarity of the noun concepts within complex descriptive phrases modulate structural preferences in production, and processing cost in comprehension. For example, animate-head phrases such as the girl that the boy is pulling are rare in production and more difficult to understand in comprehension. In contrast, phrases with passive clauses such as the girl being pulled by the boy are commonly produced and more easily understood, as are inanimate-head structures such as the truck the boy is pulling. In three picture-based studies, we examined the mechanisms underlying semantic similarity effects in producing and comprehending these phrases. Study 1 investigated structural preferences in production, whereas Study 2 investigated processing cost in comprehension. Study 3 used eye-tracking to examine the time-course of production processes. The results showed that semantic similarity elicited competition during phrase planning, influenced the choice of syntactic structure in production, and engendered comprehension difficulty in animate-head active configurations. Structural preferences, fixation probabilities reflecting production planning processes and comprehension cost significantly correlated with measures of conceptual similarity across the three studies. We argue that similarity-based competition modulates sentence production and comprehension processes when verbs are planned or interpreted, i.e., when event-based semantic or syntactic roles are determined. In addition to task-specific processes, we suggest that a similar and shared semantic competition mechanism underlies both production and comprehension, a view consistent with existing evidence for common brain regions recruited in both tasks.